Skip to main content

Full text of "A multi-space model for Chinese bids evaluation with analyzing"

See other formats


A Multi-Space Model for 

Chinese Bids Evaluation with Analyzing 

Linfan Mao 

Chinese Academy of Mathematics and System Sciences, Beijing 100080, P.R.China 
Guoxin Tendering Co.,LTD } Beijing 100044, P-R-China 

maolinfan@l 63. com 

Abstract. A tendering is a negotiating process for a contract through by 
a tenderer issuing an invitation, bidders submitting bidding documents and 
the tenderer accepting a bidding by sending out a notification of award. As 
a useful way of purchasing, there are many norms and rulers for it in the 
purchasing guides of the World Bank, the Asian Development Bank, • • •, also 
in contract conditions of various consultant associations. In China, there is 
a law and regulation system for tendering and bidding. However, few works 
on the mathematical model of a tendering and its evaluation can be found in 
publication. The main purpose of this paper is to construct a Smarandache 
multi-space model for a tendering, establish an evaluation system for bidding 
based on those ideas in the references [7] and [8] and analyze its solution by 
applying the decision approach for multiple objectives and value engineering. 
Open problems for pseudo-multi-spaces are also presented in the final section. 
Key Words: tendering, bidding, evaluation, Smarandache multi-space, 
condition of successful bidding, decision of multiple objectives, decision of 
simply objective, pseudo-multiple evaluation, pseudo-multi-space. 
AMS(20 00): 90B50,90C35,90C90 

§1. Introduction 

The tendering is an efficient way for purchasing in the market economy. According to 
the Contract Law of the People 's Republic of China (Adopted at the second meeting 
of the Standing Committee of the 9th National People's Congress on March 15,1999), 
it is just a civil business through by a tenderer issuing a tendering announcement 
or an invitation, bidders submitting bidding documents compiled on the tendering 
document and the tenderer accepting a bidding after evaluation by sending out a 
notification of award. The process of this business forms a negotiating process of a 
contract. In China, there is an interval time for the acceptation of a bidding and 
becoming effective of the contract, i.e., the bidding is accepted as the tenderer send 
out the notification of award, but the contract become effective only as the tenderer 
and the successful bidder both sign the contract. 

In the Tendering and Bidding Law of the People's Republic of China (Adopted at 
the 11th meeting of the Standing Committee of the 9th National People's Congress 
on August 30,1999), the programming and liability or obligation of the tenderer, 



1 



the bidders, the bid evaluation committee and the government administration are 
stipulated in detail step by step. According to this law, the tenderer is on the side 
of raising and formulating rulers for a tender project and the bidders are on the 
side of response each ruler of the tender. Although the bid evaluation committee is 
organized by the tenderer, its action is independent on the tenderer. In tendering 
and bidding law and regulations of China, it is said that any unit or person can 
not disturbs works of the bid evaluation committee illegally. The action of them 
should consistent with the tendering and bidding law of China and they should place 
themselves under the supervision of the government administration. 

The role of each partner can be represented by a tetrahedron such as those shown 
in Fig.l. 




Fig.l 

The 41th item in the Tendering and Bidding Law of the People's Republic of 
China provides conditions for a successful bidder: 

(1) optimally responsive all of the comprehensive criterions in the tendering 
document; 

(2) substantially responsive criterions in the tender document with the lowest 
evaluated bidding price unless it is lower than this bidder's cost. 

The conditions (1) and (2) are often called the comprehensive evaluation method 
and the lowest evaluated price method. In the same time, these conditions also 
imply that the tendering system in China is a multiple objective system, not only 
evaluating in the price, but also in the equipments, experiences, achievements, staff 
and the programme, etc.. However, nearly all the encountered evaluation methods 
in China do not apply the scientific decision of multiple objectives. In where, the 
comprehensive evaluation method is simply replaced by the 100 marks and the lowest 
evaluated price method by the lowest bidding price method. Regardless of whether 
different objectives being comparable, there also exist problems for the ability of 
bidders and specialists in the bid evaluation committee creating a false impression 
for the successful bidding price or the successful bidder. The tendering and bidding 
is badly in need of establishing a scientific evaluation system in accordance with 



2 



these laws and regulations in China. Based on the reference [7] for Smarandache 
multi-spaces and the mathematical model for the tendering in [8], the main purpose 
of this paper is to establish a multi-space model for the tendering and a scientific 
evaluation system for bids by applying the approach in the multiple objectives and 
value engineering, which enables us to find a scientific approach for tendering and 
its management in practice. Some cases are also presented in this paper. 

The terminology and notations are standard in this paper. For terminology and 
notation not defined in this paper can be seen in [7] for multi-spaces, in [1] — [3] and 
[6] for programming, decision and graphs and in [8] for the tendering and bidding 
laws and regulations in China. 

§2. A multi-space model for tendering 

Under an idea of anti-thought or paradox for mathematics -.combining different fields 
into a unifying field, Smarandache introduced the conception of multi-spaces in 
1969([9]-[12]), including algebraic multi-spaces and multi-metric spaces. The con- 
tains the well-known Smarandache geometries ([5] — [6]), which can be used to Gen- 
eral Relativity and Cosmological Physics([7]) . As an application to Social Sciences, 
multi-spaces can be also used to establish a mathematical model for tendering. 
These algebraic multi-spaces are defined in the following definition. 

Definition 2.1 An algebraic multi-space w ith multiple m is a union of m sets 
Ai, A2, • • • , A m 

m 

E = U4, 

i=i 

where 1 < m < +00 and there is an operation or ruler on each set Ai such that 
(AjOj) is an algebraic system for any integer i, 1 < % < to. 

Notice that if i ^ j, 1 < i, j < m, there must not be A^ f] Aj = 0, which are just 
correspondent with the characteristics of a tendering. Thereby, we can construct a 
Smarandache multi-space model for a tendering as follows. 

Assume there are m evaluation items A±, A 2 , • • • , A m for a tendering A and there 
are rij evaluation indexes an, a^, • • • , for each evaluation item Ai, 1 < i < m. 
By applying mathematics, this tendering can be represented by 

m 

A=\jAi, 

i=i 

where, for any integer i, 1 < % < m, 

(Ai,°i) = {«il,«i2," • - ^mJOi} 

is an algebraic system. Notice that we do not define other relations of the tendering 
A and evaluation indexes with Ai, \ < i < m unless Ai C A and a^- G Ai in this 
multi-space model. 



3 



Now assume there are k, k > 3 bidders Ri, R 2 , ■ ■ ■ , R k in the tendering A and 
the bidding of bidder Rj , 1 < j < k is 



Rj(A) = R 3 



A 2 = Rj(A 2 ) 
\A m J V Rj(A m ) J 



According to the successful bidding criterion in the Tendering and Bidding Law 
of the People's Republic of China and regulations, the bid evaluation committee 
needs to determine indexes ii, i 2 , • ■ ■ , ik, where {i±, i 2 , • ■ ■ , ik} = {1, 2, • • • , k} such 
that there is an ordered sequence 

RiM) yR i2 (A)y---y R ik (A) 

for these bidding Ri(A), R 2 (A), ■ ■ ■ , Rk{A) of bidders Ri, R 2 , ■ ■ ■ , Rk- Here, these 
bidders R^,Ri 2 and i?« 3 are pre- successful bidders in succession determined by the 
bid evaluation committee in the laws and regulations in China. 

Definition 2.2 An ordered sequence for elements in the symmetry group S n on 
{1, 2, • • • , m} is said an alphabetical sequence if it is arranged by the following crite- 
rions: 

(i) (1, • • • , 0) y P for any permutation P G S n . 

(ii) if integers s±, s 2 , • • ■ , Sh € {1, 2, • • • , m}, 1 < h < m and permutations (si, s 2 , 
■ ■ ■ , s h , t, ■ ■ •), (si, s 2 , ■ ■ ■ , s h , I, ■ ■ •) G S n , then 



(si, s 2 , ■ ■ ■ , s h ,t, ■■■)>- (si, s 2 , 



if and only ift<l. Let {x ff J" be a sequence, where cri >- a 2 >- ■ ■ ■ >- a n and a; L e S n 
for 1 < i < n, then the sequence {x^}™ is said an alphabetical sequence. 

Now if x a >- x T , x a is preferable than x T in order. If x a >_ x T , then x a is 
preferable or equal with x T in order. If x a >z x T and x T >z x a , then x a is equal x T in 
order, denoted by Xfj ~ x T . 

We get the following result for an evaluation of a tendering. 

Theorem 2.1 Let 1 , 2 , 3 ■ ■ ■ be ordered sets. If Rj(A) e O x x 2 x 3 x • • • for 

any integer j, 1 < j < k, then there exists an arrangement z l5 i 2 , • • ■ , i k for indexes 
1, 2, • • • , k such that 

R^ (A) hR i2 (A)y---y R ik (A). 

Proof By the assumption, for any integer j, 1 < j ' < k, 

Rj{A) e Oi x 2 x 3 x 



4 



Whence, Rj(A) can be represented by 

Rj(A) = (xjuXfrXja, ■ ■ •), 
where Xj t e O t , t > 1. Define a set 

St = 1 < j < m}. 

Then the set S t C O t is finite. Because the set O t is an ordered set, so there exists 
an order for elements in S t . Not loss of generality, assume the order is 

X U h X 2t h ■ ■ ■ h X mt , 

for elements in S t . Then we can apply the alphabetical approach to R il (A) , R i2 ( A) , 
• • • , R ik (A) and get indexes z'i, i 2 , ■ • • , ik such that 

R h {A)hRiM)h---hRi k {A). \ 

If we choose Oi, % > 1 to be an ordered function set in Theorem 2.1, particularly, 
let 0\ = {/}, / : Ai — > i?, 1 < z < m be a monotone function set and O t = for 
t > 2, then we get the next result. 

Theorem 2.2 Let Rj : Ai R, 1 < i < m, 1 < j < k be monotone functions. Then 
there exists an arrangement i±, i 2 , ■ ■ ■ , ik for indexes 1, 2, • • • , k such that 

R il {A)hR i2 {A)y---yR ik (A). 

We also get the following consequence for evaluation numbers by Theorem 2.2. 

Corollary 2.1 If Rj(Aj) e [— oo,+oo] x [— oo,+oo] x [— oo, +oo] x ••• for any 
integers i,j,l<i<m,l<j<k, then there exists an arrangement ii, i 2 , • ■ • , ik for 
indexes 1, 2, • • • , k such that 

R i M)hR ia (A)h'-->zRi k (A). 

Notice that in the above ordered sequence, if we arrange R ig >- R il or R it >- R is 
further in the case of Ri s ~i^,,s^l, then we can get an ordered sequence 

R h (A) y R l2 (A) y ■ ■ ■ y R lk (A), 
and the pre-successful bidders accordance with the laws and regulations in China. 

§3. A mathematical analog for bids evaluation 

For constructing an evaluation system of bids by the multi-space of tendering, the 
following two problems should be solved in the first. 



5 



Problem 1 For any integers i,j, 1 < i, j < m, how to determine Rj(Ai) on account 
of the responsiveness of a bidder Rj on indexes a tl , a i2 , ■ ■ ■ , a,i ni ? 
Problem 2 For any integer j, 1 < j < m, how to determine Rj(A) on account of 
the vector (R j (A 1 ), Rj(A 2 ), • • • , Rj(A m )Y ? 

Different approaches for solving Problems 1 and 2 enable us to get different 
mathematical analogs for bids evaluation. 

3.1. An approach of multiple objectives decision 

This approach is originated at the assumption that Rj(Ai), Rj(A 2 ), • • • , Rj(A m ), 1 < 
j < m are independent and can not compare under a unified value unit. The 
objectives of tendering is multiple, not only in the price, but also in the equipments, 
experiences, achievements, staff and the programme, etc., which are also required by 
the 41th item in the Tendering and Bidding Law of the People's Republic of China. 

According to Theorems 2.1 — 2.2 and their inference, we can establish a pro- 
gramming for arranging the order of each evaluation item A i} l < i < m and getting 

an ordered sequence of bids Ri(A), R 2 (A), ■ ■ ■ , Rk{A) of a tendering A = [j A i} as 
follows: 

STEP 1 determine the order of the evaluation items A±, A 2 , ■ • ■ , A m . For ex- 
ample, for m = 5, Ai >- A 2 pa A 3 >- A A pa A 5 is an order of the evaluation items 
A 1 ,A 2 ,A 3 ,A 4 ,A 5 . 

STEP 2 for two bids Rj^AA, Rj 2 (Ai), ji ^ j 2 , 1 < i < m, determine the con- 
dition for Rj 1 (A i ) pa Aj 2 (A 2 ). For example, let A± be the bidding price. Then 
Rj^Ai) pa Rj 2 (Ai) providing {Rj^A) — Rj 2 (Ai)\ < 100(10 thousand yuan). 

STEP 3 for any integer i, 1 < i < m, determine the order of R\{Ai) , R 2 {A,j) , 
• • • , Rk(Ai). For example, arrange the order of bidding price from lower to higher 
and the bidding programming dependent on the evaluation committee. 

STEP 4 alphabetically arrange R\(A) , R 2 (A) , • • • , Rk(A) , which need an ap- 
proach for arranging equal bids Rj 1 (A) pa Rj 2 (A) in order. For example, arrange 
them by the ruler of lower price preferable and get an ordered sequence 

R h (A) y R i2 (A) y ■ ■ ■ y R ik {A) 

of these bids Ri{A) , R 2 {A) , ■ ■ ■ , R k {A) . 

Notice that we can also get an ordered sequence through by defining the weight 
functions 

cu(A) = H{uj{A 1 ),uj{A 2 ), • • • , u(A m )) 

and 

Ld(Ai) = F(u(a a ),uj(a i2 ), ■ ■ ■ ,u(a ini )). 
For the weight function in detail, see the next section. 



6 



Theorem 3.1 The ordered sequence of bids of a tendering A can be gotten by the 
above programming. 

Proof Assume there are k bidders in this tendering. Then we can alphabetically 
arrange these bids R il (A) , R i2 (A) , • • • , R ik (A) and get 

R h {A) h Ri 2 (A) y ■■■ y R lk {A). 

Now applying the arranging approach in the case of Rj 1 (A) pa Rj 2 (A), we finally 
obtain an ordered sequence 

R il {A)yR i2 {A)y---yR ik {A). \ 

Example 3.1 There are 3 evaluation items in a building construction tendering A 
with Ai =price, A 2 =programming and A 3 =similar achievements in nearly 3 years. 
The order of the evaluation items is Ai y A 3 y A 2 and Rj^Ai) pa Rj 2 (Aj), 1 < i < 3 
providing {Rj^Ai) — Rj 2 (Ai)\ < 150, Rj^A^ and Rj 2 (A 2 ) are in the same rank or 
the difference of architectural area between Rj^A^) and Rj 2 (A 3 ) is not more than 
40000m 2 . For determining the order of bids for each evaluation item, it applies the 
rulers that from the lower to the higher for the price, from higher rank to a lower 
rank for the programming by the bid evaluation committee and from great to small 
amount for the similar achievements in nearly 3 years and arrange Rj 1 (A), Rj 2 (A), 
1 < ji, 3i < k =bidders by the ruler of lower price first for two equal bids in order 
R h {A)p*R j2 {A). 

There were 4 bidders R 1 , R 2 , R3, R4 in this tendering. Their bidding prices are 
in table 1. 



bidder 


Ri 


R2 


#3 


i?4 


A l 


3526 


3166 


3280 


3486 



tabje 1 

Applying the arrangement ruler for A\, the order for R 2 (A\) , R s (Ai) , R^(Ai) , 
Ri(Ai) is 

R 2 (A 1 ) ^ R 3 (A 1 ) y R4A ± ) pa R 1 (A 1 ). 

The evaluation order for A 2 by the bid evaluation committee is R 3 (A 2 ) pa R 2 (A 2 ) y 
-^1(^2) y Ri(A 2 ). They also found the bidding results for A 3 are in table 2. 



bidder 


Ri 


Ri 


R3 


i?4 


A 3 (m 2 ) 


250806 


210208 


290108 


300105 



tabje 2 

Whence the order of R 4 (A 3 ) , R 3 (A 3 ) , R X (A 3 ) , R 2 (A 3 ) is 



7 



^4(^3) « R 3 (A 3 ) y R 1 (A 3 ) « i? 2 (^ 3 )- 
Therefore, the ordered sequence for these bids Ri(A), R 2 (A), R 3 (A) and R^A) is 

R 3 (A) y R 2 (A) y R 4 (A) y R 1 (A). 

Let the order of evaluation items be A\ y A 2 y ■ ■ ■ y A m . Then we can also 
get the ordered sequence of a tendering by applying a graphic method. By the 
terminology in graph theory, to arrange these bids of a tendering is equivalent to 
find a directed path passing through all bidders Ri, R 2 , ■ ■ ■ , Rk in a graph G[A] 
defined in the next definition. Generally, the graphic method is more convenience in 
the case of less bidders, for instance 7 bidders for a building construction tendering 
in China. 

~ m 

Definition 3.1 Let Ri, R 2 , ■ ■ ■ , Rk be all these k bidders in a tendering A = [j A\. 

i=i 

Define a directed graph G[A] = (V{G[A\), E(G[A})) as follows. 

V(G[A\) = {R 1 ,R 2 ,---, R k } x{A u A 2 ,---, A m }, 

E(G[A}) = E 1 {JE 2 {JE 3 . 
Where E 1 consists of all these directed edges (R^^Ai) , Rj 2 (Ai)) , 1 < i < m, 1 < 
ji)j2 < k and Rj^Aj) y Rj 2 (A{) is an adjacent order. Notice that if R s (Ai) m 
Ri(Ai) y Rj(Ai), then there are R s {Ai) y Rj(Ai) and Ri(Ai) y Rj(Ai) simulta- 
neously. E 2 consists of edges Rj 1 (Ai)Rj 2 (Ai),l < i < m, 1 < ji,j 2 < k, where 
R h {Ai) m R h (Ai) and E 3 = {R j {A i )R j (A i+1 )\l < % < m - 1, 1 < j < k}. 

For example, the graph G[A] for Example 3.1 is shown in Fig. 2. 




Fig.2 



8 



Now we need to find a directed path passing through Ri, R 2 , -R3, R4 with start 
vertex R 2 (Ai) or R 3 (Ai). By the ruler in an alphabetical order, we should travel 
starting from the vertex R^(Ai) passing through A 2 ,A% and then arriving at A\. 
Whence, we find a direct path correspondent with the ordered sequence 

R 3 (A) y R 2 (A) y R 4 (A) y R 1 (A). 
3.2. An approach of simply objective decision 

This approach is established under the following considerations for Problems 1 and 
2. 

Consideration 1 In these evaluation items A±, A 2 , • ■ ■ , A m of a tendering A, seek 
the optimum of one evaluation item. For example, seek the lowest bidding price in a 
construction tendering for a simply building or seek the optimum of design scheme 
in a design project tendering, etc.. 

Consideration 2 The value of these evaluation items A±, A 2 , ■ ■ ■ , A m is comparable 
which enables us to measure each of them by a unify unit and to construct various 
weighted functions on them. For example, the 100 marks and the lowest evaluated 
price method widely used in China are used under this consideration. 

3.2.1. The optimum of one objective 

Assume the optimal objective being Ai in a tendering A = U Aj. We need to deter- 

i=\ 

mine the acceptable basic criterions for all other items A 2 , A 3 , ■ ■ ■ , A&, then arrange 
Ri(Ai), R 2 (A 1 ),•••, Ri(Ai) among these acceptable bids R±, R 2 , ■ ■ ■ , Ri for items 
A 2 , A 3 , ■ ■ ■ , Ak in 1 < i < k. For example, evaluating these items A 2 , A 3 , • • • , A& 
by qualification or by weighted function on A 2 , A 3 , • • • , A& up to these criterions, 
then arrange these acceptable bids R±,R 2 ,---,Ri under their response to A 1 and 
the order of Ri(A), Ri(A) if Ri(Ai) R j (A 1 ). According to Theorem 3.1, we get 
the following result. 

Theorem 3.2 The approach of one optimal objective can get an ordered sequence 
of bids for a tendering A. 

Example 3.2 The optimum of design scheme is the objective in a design project 
tendering A which is divided into 5 ranks A, B, C, D, E and other evaluation items 
such as human resources, design period and bidding price by a qualifiable approach 
if the bidding price is in the interval of the service fee norm of China. The final 
order of bids is determined by the order of design schemes with qualifiable human 
resources, design period and bidding price and applying the ruler of lower price first 
for two equal design scheme in order. 

There were 8 bidders in this tendering. Their bidding prices are in table 3. 



bidder 


Ri 


R2 


R3 


i?4 


#5 


R& 


Rr 


Rs 


bidding price 


251 


304 


268 


265 


272 


283 


278 


296 



9 



tabje 3 

After evaluation for these human resources, design period and bidding price, 4 
bidders are qualifiable unless the bidder R 5 in human resources. The evaluation 
result for bidding design schemes is in table 4. 



rank 


A 


B 


C 


D 


E 


design scheme 


R^R% 


Ri 


R2R9, 


R7 


i?4 



tabje 4 

Therefore, the ordered sequence for bids is 

R 3 (A) y R 6 (A) y R^A) y R 8 (A) y R 2 (A) y R 7 (A) y R 4 (A). 

Example 3.3 The optimum objective in a tendering A for a construction of a 
dwelling house is the bidding price A\. All other evaluation items, such as qualifica- 
tions, management persons and equipments is evaluated by a qualifiable approach. 

There were 7 bidders 1 < i < 7 in this tendering. The evaluation of price 
is by a weighted function approach, i.e., determine the standard price S first, then 
calculate the mark N of each bidder by the following formulae 

(E Ai - max{i^(Ai)|l < i < 7} - min-fi^Ai)!! < i < 7} 



N i = 100- tx\^^ -|xlOO, l<t<7, 

o 

where, if Ri(A\) — S > then t = 6 and if Ri(A\) — S < then t = 3. 

After evaluation, all bidders are qualifiable in qualifications, management persons 
and equipments. Their bidding prices are in table 5. 



bidder 


Ri 


R2 


#3 


i?4 


R 5 


#6 


R7 


A 1 


3518 


3448 


3682 


3652 


3490 


3731 


3436 



tabje 5 

According to these formulae, we get that S = 3558 and the mark of each bidder 
as those shown in table 6. 



bidder 


Ri 


R2 


Rs 


i?4 


R 5 


Rq 


R 7 


mark 


96.70 


91.27 


79.12 


84.16 


94.27 


73.84 


89.68 



10 



tabje 6 

Therefore, the ordered sequence of bids is 

R^A) y R 5 (A) y R 2 (A) y R 7 (A) y R A (A) y R 3 (A) y R 6 (A). 
3.2.2. The pseudo-optimum of multiple objectives 

This approach assumes that there is a unifying unit between these evaluation items 
Ai, A 2 , • • • , A m in an interval [a, b]. Whence it can be transformed into case 3.2.1 
and sought the optimum of one objective. Not loss of generality, we assume the 
unifying unit is w and 

w(Ai) = fi(w), 1 < % < m, 

where fi denotes the functional relation of the metric w{Aj) with unit w. Now the 
objective of tendering turns to a programming of one objective 

maxF(/i(tz7), f 2 {w),---, f m {w)) or mm F(f 1 {w), f 2 {w), ■■■,f m {w)), 

where F denotes the functional relation of the tendering A with these evaluation 
items A±, A 2 , • • • , A m , which can be a weighted function, such as a linear function 

m 

F(fi(ru),f 2 (w),---,f m (w)) = J2M^) 

i=l 

or an ordered sequence. According to Theorem 3.2, we know the following result. 

Theorem 3.3 // the function F of a tendering A only has one maximum value in 
[a,b], then there exists an ordered sequence for these bids Ri(A),l < i < k after 
determined how to arrange Ri(A) and Rj(A) when F(Ri(A)) = F(Rj(A)),i ^ j. 

The 100 marks and the lowest evaluated price method widely used in China both 
are applications of this approach. In the 100 marks, the weight function is a linear 
function 

m 

F(fi(uj), f 2 (m), ■ ■ ■ , f m (w)) = ^2fi(w) 

i=i 

with < F{f l {w),f 2 {w) r -- ,f m {w)) < 100, /i > 0, 1 < i < m. In the lowest 
evaluated price method, each difference of an evaluation item Ai,2 < i < m is 
changed to the bidding price zu(Ai), i.e., 

fi = - 1 < % < m, 

where S(Ai) is the standard line for Ai, zu(Ai) is one unit difference of Ai in terms 
of Ai. The weighted function of the lowest evaluated price method is 



11 



F(w(A 1 ), f 2 {zu(A 1 )), fm^A,))) = (1 + ^(A) - S(4))M^i). 

i=2 

For example, we can fix one unit difference of a technological parameter 15, i.e., 
va{A 1 ) = 15 ten thousand dollars in terms of the bidding price. 

§4. Weighted functions and their construction 

We discuss weighted functions on the evaluation items or indexes in this section. 
First, we give a formal definition for weighted functions. 

Definition 4.1 For a tendering A = [j A it where Ai = {an, a i2 , ■ ■ ■ , aj„}, 1 < i < m 

i=i 

with k bidders R 1 , R 2 , • ■ • , R^, if there is a continuous function uj : A — > [a, b] C 
(— oo, +oo) or uj : Ai — > [a, b] C (— oo, +oo), 1 < i < m such that for any integers 
l,s,l < l,s < k,_Ri(u(A)) > R s (lo(A)) or Ri(uj(A)) = R s (uj(A)) as Ri(A) y R S (A) 
orR^A) w R S {A) and Ri{uj{Ai) > R s (u(Ai)) or Ri(cu(Ai)) = R a (u(Ai)) as Ri{A-) y 
R s (Ai) or Ri(Ai) ps R s (Ai), 1 < % < m, then uj is called a weighted function for the 
tendering A or the evaluation items Ai, 1 < i < m. 

According to the decision theory of multiple objectives([3]), the weighted function 
cj(Ai) must exists for any integer i, 1 < % < m. but generally, the weight function 
oj(A) does not exist if the values of these evaluation items Ai, A 2 , • • • , A m can not 
compare. There are two choice for the weighted function cu(Ai). 

Choice 1 the monotone functions in the interval [a, b], such as the linear functions. 

Choice 2 The continuous functions only with one maximum value in the interval 
[a, b], such as uj(Ai) = —2x 2 + 6x + 12 or 



cu(Ai) = 




if < x < 2, 
if x > 4. 



As examples of concrete weighted functions uj, we discuss the tendering of civil 
engineering constructions. 

4.1. The weighted function for the bidding price 

Let Ai be the bidding price. We often encounter the following weighted function 
uj(Ai) in practice. 

RiUJ-S A 
u(R i (A 1 )) = -q X + C 

where, 



12 



R^) + R 2 (A 1 ) + ■ ■ ■ + R^) 



k 

or 



( Ri(A 1 )+R 2 (A 1 )+---+R k (A 1 )-M-: - , r 
C _ J fc-2 ' K - °' 

° | Ri(Ai)+fi 2 (Ai)+-+fi fc (Ai) 3 < A; < 4 



or 



S = TxA% + R ^ + R ^ + - + R ^ x(l-A%). 

k 

Where T : A% : k, M and TV are the pre-price of the tender, the percentage of T in S, 
the number of bidders and the maximum and minimum bidding price, respectively, 
Ri(Ai), i = 1, 2, • • • , k denote the bidding prices and <;, £ are both constants. 

There is a postulate in these weighted functions, i.e., each bidding price is random 
and accord with the normal distribution. Then the best excepted value of this civil 
engineering is the arithmetic mean of these bidding prices. However, each bidding 
price is not random in fact. It reflects the bidder's expected value and subjectivity 
in a tendering. We can not apply any definite mathematics to fix its real value. 
Therefore, this formula for a weighted function can be only seen as a game, not a 
scientific decision. 

By the view of scientific decision, we can apply weighted functions according to 
the expected value and its cost in the market, such as 

(1) the linear function 

Ri(Ai) - N 
W)) = -F tK M ^ N +q 

in the interval [N, M] , where M, N are the maximum and minimum bidding prices 
p is the deduction constant and q is a constant such that Ri(u(Ai)) > 0, 1 < i < k. 
The objective of this approach is seek a lower bidding price. 

(2) non-linear functions in the interval [N, M], such as 

k 



w fii A =-px v + q 

k ^R 1 (A 1 )R 2 (A 1 )---R k (A 1 )T 



or 



13 



/ Rl(A 1 )+Rl(A 1 )+-+Rl(A 1 )+T2 
V k+1 

etc.. If we wish to analog a curve for these bidding prices and choose a point on this 
curve as u(Ri(Ai)), we can apply the value of a polynomial of degree k + 1 

f(x) = a k+1 x k+1 + a k x k H h a±x + o 

by the undetermined coefficient method. Arrange the bidding prices and pre-price 
of the tender from lower to higher. Not loss of generality, let it be Rj 1 (A 1 ) y 
Rp 2 )(A 1 ) y ■ ■ • y T y ■ ■ • y R jk (A 1 ). Choose k + 2 constants ci > c 2 > ■ • • > 
Cfe + i > 0, for instance A; + 1>A;>--->1>0. Solving the equation system 

Rj^Ai) = a k+ ic k+1 + a k c k H h a x c x + a 

Rj 2 (Ai) = a k+ iC2 +l + a k c\ H h aic 2 + a 



fl^^Ai) = a fc+ iCfc +1 + a k c k k H h a x c k + o 

we get a polynomial f(x) of degree A; + 1. The bidding price has an acceptable 
difference in practice. Whence, we also need to provide a bound for the difference 
which does not affect the ordered sequence of bids. 

4.2. The weighted function for the programming 

Let A 2 be the evaluation item of programming with evaluation indexes {021,0221 
• • • , «2n 2 }- K is difficult to evaluating a programming in quantify, which is not only 
for the tender, but also for the evaluation specialists. In general, any two indexes of 
A 2 are not comparable. Whence it is not scientific assigning numbers for each index 
since we can not explain why the mark of a programming is 96 but another is 88. 
This means that it should qualitatively evaluate a programming or a quantify after 
a qualitatively evaluation. Its weight function cu(Ri(A 2 )), 1 < i < k can be chosen 
as a linear function 

u(Ri(A 2 )) = u{Ri{a 21 )) + io(Ri(a 22 )) + ■■■ + u(R t (a 2n2 )). 

For example, there are 4 evaluation indexes for the programming, and each with 
A, B,C,D ranks in a tendering. The corespondent mark for each rank is in table 7. 



index 


a 2 \ 


a 22 


«23 


a 2 4 


A 


4 


2 


2 


1 


B 


3 


1.5 


1.5 


0.8 


C 


2 


1 


1 


0.5 


D 


1 


0.5 


0.5 


0.3 



14 



tabje 7 

If the evaluation results for a bidding programming R iy 1 < i < 4 are uj(Ri(a 2 i)) = 
A, u(Ri(a 22 )) = B, uj(Ri(a 23 )) = B and u(Ri(a 2 4)) = A, then the mark of this pro- 
gramming is 

Ri(u(A 2 )) = Ri(u(a 21 )) + Ri{uj{a 22 )) + Ri(u;(a 23 )) + Ri(u(a 24 )) 
= 4 + 3 + 1.5+1 = 9.5. 

By the approach in Section 3, we can alphabetically or graphicly arrange the 
order of these programming if we can determine the rank of each programming. 
Certainly, we need the order of these indexes for a programming first. The index 
order for programming is different for different constructions tendering. 

§5. Further discussions 

5.1 Let A = U Ai be a Smarandache multi-space with an operation set 0(A) = 

i=i 

{oj; 1 < i < m}. If there is a mapping 6 on A such that 0(A) is also a Smarandache 
multi-space, then (A, O) is called a pseudo-multi-space. Today, nearly all geometries, 
such as the Riemann geometry, Finsler geometry and these pseudo-manifold geome- 
tries are particular cases of pseudo-multi-geometries. 

For applying Smarandache multi-spaces to an evaluation system, choose 0(A) 
being an order set. Then Theorem 3.1 only asserts that any subset of 0(A) is an 
order set, which enables us to find the ordered sequence for all bids in a tendering. 
Particularly, if 0(A) is continuous and 0(A) C [— oo,+oo], then O is a weighted 
function on A widely applied in the evaluation of bids in China. By a mathematical 
view, many problems on (A, O) is valuable to research. Some open problems are 
presented in the following. 

Problem 5.1 Characterize these pseudo-multi- spaces (A, O), particularly, for these 
cases of 0(A) = U [a*, h\, 0(A) = U (G h o,) and 0(A) = [j (R- + h 0i ) with (G h oA 

i=l i=l i=l 

and (R; +j, Oj) being a finite group or a ring for 1 < i < n. 

Problem 5.2 Let 0(A) be a group, a ring or a filed. Can we find an ordered sequence 
for a finite subset of A? 

Problem 5.3 Let 0(A) be n lines or n planes in an Euclid space R™. Characterize 
these pseudo-multi- spaces (A, O). Can we find an arrangement for a finite subset of 
A? 

5.2 The evaluation approach in this paper can be also applied to evaluate any multi- 
ple objectives, such as the evaluation of a scientific project, a personal management 
system, an investment of a project, • • •, etc.. 



15 



References 



G.Chartrand and L.Lesniak, Graphs & Digraphs, Wadsworth, Inc., California, 
1986. 

T.Chen, Decision Approaches in Multiple Objectives(in Chinese), in The Hand- 
book of Modern Engineering Mathematics, Vol.IV, Part 77,1357-1410. Central 
China Engineering College Press, 1987. 

P.C.Fishburn, Utility Theory for Decision Making, New York, Wiley, 1970. 
D.L.Lu, X.S.Zhang and Y.Y.Mi, An offer model for civil engineering construc- 
tion, Chinese OR Transaction, Vol.5, No.4(2001)41-52. 

L.F.Mao, On Automorphisms groups of Maps, Surfaces and Smarandache ge- 
ometries, Sientia Magna, Vol. 1(2005), No. 2, 55-73. 

L.F.Mao, Automorphism Groups of Maps, Surfaces and Smarandache Geome- 
tries, American Research Press, 2005. 

L.F.Mao, Smarandache multi-space theory, Hexis, Phoenix, AZ2006. 
L.F.Mao, Chinese Construction Project Bidding Technique & Cases Analyzing- 
Smarandache Multi-Space Model of Bidding,Xiquan Publishing House (Chinese 
Branch), America, 2006. 

F. Smarandache, Mixed noneuclidean geometries, eprint arXiv: math/0010119, 
10/2000. 

F. Smarandache, A Unifying Field in Logics. Neutrosopy: Neturosophic Proba- 
bility, Set, and Logic, American research Press, Rehoboth, 1999. 
F. Smarandache, Neutrosophy, a new Branch of Philosophy, Multi-Valued Logic, 
Vol.8, No. 3 (2002) (special issue on Neutrosophy and Neutrosophic Logic), 297- 
384. 

F. Smarandache, A Unifying Field in Logic: Neutrosophic Field, Multi-Valued 
Logic, Vol.8, No. 3(2002) (special issue on Neutrosophy and Neutrosophic Logic), 
385-438. 



16